CN1336769A - Image code having movement predicting section - Google Patents
Image code having movement predicting section Download PDFInfo
- Publication number
- CN1336769A CN1336769A CN01119334A CN01119334A CN1336769A CN 1336769 A CN1336769 A CN 1336769A CN 01119334 A CN01119334 A CN 01119334A CN 01119334 A CN01119334 A CN 01119334A CN 1336769 A CN1336769 A CN 1336769A
- Authority
- CN
- China
- Prior art keywords
- section
- image
- macro block
- motion
- estimation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/55—Motion estimation with spatial constraints, e.g. at image or region borders
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention relates to a picture coding method for transmission on a channel at a low bit rate and with a high error rate, typically a channel with a bit rate less than 100 kbit/s and an error rate greater than 10-6, or even 10-4. The method uses motion estimation coding and divides the picture into a plurality of segments made up of macroblocks. According to the invention, at least for an edge macroblock of a segment, a motion estimation vector is allowed to extend into an adjacent segment. This is particularly advantageous for macroblocks situated top right of the segment if scanning for motion vector estimation is from left to right and top to bottom. This solution avoids errors caused by estimating motion vectors within segments and limits the propagation of coding errors caused by splitting the picture into segments.
Description
The present invention relates to coding, especially for the source video coding of wireless transmission.
Wireless transmission is a feature with relatively low transmission rate and higher error rate.For transmission rate, we are lower than 100kbit/s (kilobits/second) to rate value and call low rate, for example the GSM of 14kbit/s (Global Link) channel.We are greater than 10
-6, or greater than 10
-4Error rate be called high error rate.
ITU-T H263+ standard provides solution for video frequency source coding.This standard is made of a compulsory public part and an optional appendix set, and the appendix set can be activated outside the public part of standard.Standard does not indicate the quality of the image that is obtained, and does not indicate the appendix combination that will use in given environment, particularly for wireless transmission yet.The startup of this appendix set causes system to be difficult to be used for wireless transmission owing to low available transmission rate and high error rate.
Standard public part has H.263+ proposed to have the block encoding of prediction.We distinguish the image of the image-type " I " of complete transmission in an image sequence, intactly do not transmitted but from the image of a front predicted-image of type " P "-and intactly do not transmitted but predicted from a previous and back image.For an image, wherein defined piece or, be typically 6 pieces by a plurality of macro blocks that constitute, 2 colourities of 4 brightness wherein, coding uses the following step:
Each image block of-estimation or the motion of determining greatly;
-carry out motion compensated prediction with respect to a reference picture;
-coding is used for transmission, is typically by compressed encoding (comprise DCT (" Discrete Cosine Transform (discrete cosine transform) "), quantize, reach VLC coding (" Variable Length Coding (variable length code) ")).
Decoding uses opposite step with each width of cloth image of reconstruct.
In this standard, appendix R-Independent Segment decoding mode or independent segment decoding schema-suggestion are used video image section (" videopicture segments ") for motion estimation process, and it is made up of a set of blocks; Motion estimation is carried out in each section independently.In the case, for decoding, the processed image of doing of the restriction of a section limits, and also comprises the processing of the moving image vector that exceeds these restrictions.These moving images that exceed restriction are at appendix alternative mode D, and F is not when J or O exist and be under an embargo.If we use the motion vector pattern (" Unrestricted motionvector mode ") of hard-core appendix D, the current section boundary of video image is extrapolated to constitute the prediction of pixel outside the section zone.
The appendix D of standard advises an exception in the principle that public part proposed of standard, promptly motion vector is limited so that all pixels of its institute's reference all are in the image encoding district.More specifically, this name is called the appendix suggestion of " Unrestricted Motion Vector Mode " (unrestricted motion vector pattern), and motion vector can point to outside the image.When a pixel by the reference of motion vector institute was outside the coded image surface, we used an edge pixel in its position.The value of edge pixel is obtained by motion vector being limited in corresponding position of complete pixel of coded image interior surface.Be performed on the basis that is limited in a pixel of motion vector, and used respectively for each component of motion pixel.
Other appendix suggestion, motion vector can pass piece image or piece image section boundary.This is appendix F (" Advanced Prediction Mode " advanced prediction mode), J (filtered model of " Deblocking filter mode " piecemeal), O (" temporal, SNR and Spatial Scalability ", time scale, signal to noise ratio or space scale) situation.
Invention has proposed a solution at low rate and high error rate transmission problem.At this effect problem of block encoding and motion vector restriction has been proposed especially above.This problem is suggested in sharp-pointed especially mode in the transmission of low rate and high error rate, and the mistake that is occurred is difficult to proofread and correct by transmission or coding adjuster again, and its reason is the delay that it produces.In addition, low rate can not be compensated by a littler number of errors.
This problem is suggested in special mode for macro block, and macro block is positioned on the right side and top edge of a section; This causes the application of encoding in advance, and particularly on the scanning direction of macro block: scan image from left to right and from high to low in fact normally is used for determining of motion vector.
What advised in appendix R in addition, is the influence that has limited transmission error by one of effect of section coding.From this point, appendix R is by in the strict especially transmission situation that is applied in low rate, and it can to come from the appendix R proposal of lower rate transmissions situation of segment limit opposite with being used for a motion vector.
Therefore invention has proposed a kind of method for encoding images that is used in the channel of low rate and high error rate, in the method, image is separated into a plurality of sections that are made of macro block, and encode by motion estimation, it is characterized in that for a macro block on the edge of a section at least one motion estimation vector can be stretched in the adjacent section.
In one embodiment, for a macro block on the edge of a section, a motion estimation vector can be stretched in the adjacent macroblocks of adjacent segment.
Best, the image scanning that is used for motion vector estimation is performed from high to low from a left side to back reaching, and for the macro block adjacent to a section top edge, a motion estimation vector can be stretched in the adjacent section.
If image scanning is from left to right to reach from high to low to be performed for the estimation of motion vector, for the macro block of a contiguous section right hand edge, the motion estimation vector preferably can be stretched in the adjacent section.
In one embodiment, each section is made of two row macro blocks.
When error rate greater than 10
-6Or greater than 10
-4, or when transmission rate during less than 100kb/s, invention particularly suitable.
Invention also proposes a kind of encoder of using that this coding method is dropped into.It also provides a transmission system that comprises this encoder on the channel of low rate and high error rate.
The feature and advantage of inventing other will display by the description according to inventive embodiments, and embodiment provides as an example and with reference to accompanying drawing, and accompanying drawing shows:
-Fig. 1, the letter of the definition that video image section is possible in piece image is shown;
-Fig. 2, one shows according to the letter of encoder of invention.
Invention is used to the transmission system of motion vector coding; Its system-these term of being specially adapted to low rate and high error rate is defined in the above more accurately.Under the situation that image sections is determined for motion vector estimation, for the piece or the macro block of definition image edge, invention proposes to accept motion vector and overflow the border on adjacent segment.
Fig. 1 shows the possible definition of a video image section in piece image.In example, image is that a QCIF image and each section are made of two row macro blocks, and limits synchronously by one.Each macro block is made of 4 luminance block and 2 chrominance block, and each piece itself takes advantage of the square of 8 pixels to constitute by one 8 pixel.Not only be used to the section of cutting apart synchronously, but also can make decoder find when on a certain section, transmission error occurring synchronously.Cutting apart of section avoided mistake at all images vertical spread.
In example, each section is made of two row macro blocks, MBi, and they are made GOBi by note, and this abbreviation is the abbreviation of " piece group " in the English.Every row is made of 11 macro blocks, and image is cut into 9 piece groups, that is to say corresponding 4 and half section of one 176 image of taking advantage of 144 luminance pixels.As shown in the figure, each section is by a section that sync splits is adjacent.
According to the appendix R of the standard H263+ that discussed in the above, the estimation of motion vector is carried out in the inside of section.This solution has the advantage that the restriction mistake spreads, and each section is encoded independently with adjacent segment.This solution is gratifying, but is limited in the transmission of low rate and high error rate.
For the motion vector estimation of the macro block at the edge of a section, invention proposes to loosen this restriction.More particularly, motion vector estimation by the situation that from left to right and from high to low scan image is carried out under, just the problem of motion estimation has been proposed for the macro block that forms each section top edge and right hand edge.In fact, for motion estimation, we use and to be in the left side, above and the going up of current pixel, right pixel; Estimation at intersegmental part is to be performed in correct mode for the macro block on the section of the being positioned at left side or the section lower edge (for those sections more than two row macro blocks).Yet for the macro block on the right hand edge of the top edge of the section of being positioned at or section, the restriction of motion vector has proposed problem.Invention proposes, and can have an end in one of adjacent macroblocks of adjacent segment.
In illustrated embodiment, we consider the section that the row by piece GOB4 and GOB5 constitutes.This section by one in the drawings note do Sync4 synchronously in the above part be limited, by be labeled as on the figure one of Sync5 synchronously in its lower section part be limited.We consider the macro block MBx of the row of macro block GOB4.One of edge macro block of this macro block section of being.Therefore for the pixel of this macro block or the motion vector estimation of piece, invention proposes not only to use other macro block of section, and uses the macro block of the capable formation section of macro block GOB2 and GOB3 where necessary.More specifically, preferably use neighboring macro-blocks uniquely, hold the macro block that capable GOB3 talked about in sentence under the arm, or more specifically, macro block MBx-1, MBx and the MBx+1 of row GOB3.
Fig. 1 shows capable jloW, and it has cut apart the row of macro block GOB2 and GOB3, and line jhigh, and it has cut apart the row of macro block GOB5 and GOB6.In this case, for the motion estimation of a macro block of section top edge, that is to say a macro block for row GOB4, we can use the macro block of being expert between jlow and the row jhigh.Distance between row jlow and the synchronous Sync4 is 15 pixels.
Invention is applied to be positioned at a macro block on the section right side especially, and this is always in image scanning from left to right under from high to low the situation.In this case, if do not allow the top edge of the motion vector section of passing, it will have the trend of horizontal stretching, and this will have the effect that produces image output.Invention can be avoided this problem.Therefore preferably invention is applied to above the section of being positioned at and the macro block on the right side, in the example of Fig. 1 be by row GOB4 and GOB5 constitute section the macro block MB10 of the capable GOB4 of piece.
For the transmission of low rate and high error rate, invention can avoid because restriction the produce mistake of motion estimation on a section.It has improved the quality of reconstructed image, and produce to disturb in the restriction that can not run in the transmission of low rate and high error rate, i.e. transmission speed, the complexity of encoder or decoder and memory capacity.Yet under the confined situation of motion vector estimation, invention has the advantage of appendix R, that is to say that it can limit the ratio of error.
In the description in front, we do not have what clear and definite reference picture is; It can be selected in known manner, and can be as the image in the encoder example of Fig. 2 simply.
Fig. 2 shows a kind of will the invention briefly and drops into the examples of encoder of using.Video image is applied to the input of coded system, and it arrives moving image estimation device 2, the input of arriving subtracter 4 on the other hand on the one hand.It provides a picture block with respect to last image motion estimation, or the macro block of this image, motion estimation; As what mention in standard, we can use the piece of 8 * 8 pixels, and 6 macro blocks that piece constitutes, wherein 4 luminance block and two chrominance block.
According to invention, motion estimator uses as the image sections of being mentioned among the appendix R.Therefore motion vector is estimated in section.Yet, as explained earlier, at least for edge block, the estimation device use can the section of exceeding motion vector.
The motion vector of piece is sent in the fallout predictor with motion compensation 6.This fallout predictor also receives the input of a basic image that is used to estimate, it can have a reference picture store 8.The prediction that has motion compensation is performed with respect to reference picture.
From reference picture and motion vector, the fallout predictor that has motion compensation provides the image of an estimation, and it is applied to the subtraction input of subtracter 4 on the one hand, is applied to the input of adder 10 on the other hand.
Like this, subtracter 4 receives video image and image of estimating from the reference picture that is contained in the basic video memory of input.The estimation image is applied to being typically a digital cosine transcriber on the encoder 12 that has a compression with respect to the difference on the video image that acts in the input.Next we quantize 14, preferably have a variable quantization step-length.The output of quantizer 14 is applied to a variable length encoder 16 on the one hand, is applied to a quantification reverser 18 on the other hand, and it quantizes reverse operating.In the output that quantizes reverser 18, the conversion of our applying encoder 12 reverse 20.In the output of coding reverser 20, we find the difference between input picture and the estimation image, and for example it when transmission error occurring, is configured after receiving in a relevant decode system.This difference is applied to the input of adder 10.Like this, adder 10 provides an image at it in the output of reference picture store, and it is obtained that this image can receive the back at decode system,--be in the case of necessary after the disturbance that transmission error causes.
In other words the output of variable length encoder received code device 14, and the output of motion estimator 2 are difference and the motion vectors between input picture and the estimation image.The output of VLC encoder was applied to a video buffer 22 before channel; Certainly, before transmission, we can carry out vision signal and audio signal or other signal multiplexed, and this itself is known and does not influence the function of invention.For a variable quantization step-length, the information quantization in buffering can be controlled quantization step, shown in the arrow that passes through a connection buffering 22 and quantizer 14 on Fig. 2.
With respect to the encoder of prior art, the difference of invention encoder is that motion estimator also considered the macro block outside the present segment, for example the macro block in the adjacent macroblocks row.We will notice, the decoder that is used for the signal that the decoding and coding device provided does not have different with the decoder of background technology.
Certainly, the present invention is not limited to embodiment and the Implementation Modes that institute describes and shows, the application that it also has numerous persons of ordinary skill in the field to predict.Very clear, the definition of macro block that provides previously or section only is an example, and invents the definition that also can be used for other section of image, or the definition of other piece and macro block.We can also be applicable to is cut into vertical section with image, and in this case, the scanning direction should change according to the applied macro block of invention fully.
Claims (9)
1. method for encoding images that is used in the channel of low rate and high error rate, in the method, image is separated into a plurality of sections that are made of macro block, and encode by motion estimation, it is characterized in that for a macro block on the edge of a section at least one motion estimation vector can be stretched in the adjacent section.
2. method according to claim 1 is characterized in that a motion estimation vector can be stretched in the adjacent macroblocks of adjacent segment for one on the edge of a section grand determining.
3. method according to claim 1 and 2, the image scanning that it is characterized in that being used for motion vector estimation from left to right reaches from high to low and is performed, and for the macro block adjacent to a section top edge, a motion estimation vector can be stretched in the adjacent section.
4. according to claim 1,2 or 3 described methods, it is characterized in that the image scanning that is used for motion vector estimation is performed from high to low from a left side to back reaching, and for the macro block adjacent to a section right hand edge, a motion estimation vector can be stretched in the adjacent section.
5. according to the described method of one of claim 1 to 4, it is characterized in that each section is made of the line of two macro blocks.
6. according to the described method of one of claim 1 to 5, it is characterized in that error rate is greater than 10
-6, even greater than 10
-4
7. according to the described method of one of claim 1 to 6, it is characterized in that transmission rate is less than 100kb/s.
8. encoder that the method input of one of claim 1 to 7 is used.
9. transmission system on the channel of low rate and high error rate that comprises the encoder of claim 8.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR0006954 | 2000-05-30 | ||
| FR0006954A FR2809917B1 (en) | 2000-05-30 | 2000-05-30 | SEGMENT IMAGE CODING WITH MOTION PREDICTION |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1336769A true CN1336769A (en) | 2002-02-20 |
| CN1167277C CN1167277C (en) | 2004-09-15 |
Family
ID=8850812
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB011193344A Expired - Fee Related CN1167277C (en) | 2000-05-30 | 2001-05-30 | Image code having movement predicting section |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US7277487B2 (en) |
| EP (1) | EP1161099A1 (en) |
| JP (1) | JP2001346217A (en) |
| CN (1) | CN1167277C (en) |
| FR (1) | FR2809917B1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100418465C (en) * | 2001-06-29 | 2008-09-17 | Bsh博施及西门子家用器具有限公司 | Vacuum cleaner |
| CN103385004A (en) * | 2011-06-30 | 2013-11-06 | 三菱电机株式会社 | Image encoding apparatus, image decoding apparatus, image encoding method and image decoding method |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6987866B2 (en) * | 2001-06-05 | 2006-01-17 | Micron Technology, Inc. | Multi-modal motion estimation for video sequences |
| WO2009037726A1 (en) * | 2007-09-18 | 2009-03-26 | Fujitsu Limited | Moving image coder and moving image decoder |
| KR100949917B1 (en) * | 2008-05-28 | 2010-03-30 | 한국산업기술대학교산학협력단 | Fast Encoding Method and System Via Adaptive Intra Prediction |
| JP5760953B2 (en) * | 2011-10-31 | 2015-08-12 | 富士通株式会社 | Moving picture decoding apparatus, moving picture encoding apparatus, moving picture decoding method, and moving picture encoding method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2231749B (en) * | 1989-04-27 | 1993-09-29 | Sony Corp | Motion dependent video signal processing |
| GB2231747B (en) * | 1989-04-27 | 1993-07-07 | Sony Corp | Motion dependent video signal processing |
| CA2079434A1 (en) * | 1991-09-30 | 1993-03-31 | Derek Andrew | Motion vector estimation, motion picture encoding and storage |
| GB2266023B (en) * | 1992-03-31 | 1995-09-06 | Sony Broadcast & Communication | Motion dependent video signal processing |
| KR100212559B1 (en) * | 1996-03-22 | 1999-08-02 | 전주범 | The contour coding system and motion prediction method for object |
| US6141448A (en) * | 1997-04-21 | 2000-10-31 | Hewlett-Packard | Low-complexity error-resilient coder using a block-based standard |
| GB2347038A (en) * | 1999-02-18 | 2000-08-23 | Nokia Mobile Phones Ltd | A video codec using re-transmission |
-
2000
- 2000-05-30 FR FR0006954A patent/FR2809917B1/en not_active Expired - Fee Related
-
2001
- 2001-05-14 EP EP01401227A patent/EP1161099A1/en not_active Ceased
- 2001-05-29 JP JP2001160695A patent/JP2001346217A/en active Pending
- 2001-05-29 US US09/865,733 patent/US7277487B2/en not_active Expired - Fee Related
- 2001-05-30 CN CNB011193344A patent/CN1167277C/en not_active Expired - Fee Related
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100418465C (en) * | 2001-06-29 | 2008-09-17 | Bsh博施及西门子家用器具有限公司 | Vacuum cleaner |
| CN103385004A (en) * | 2011-06-30 | 2013-11-06 | 三菱电机株式会社 | Image encoding apparatus, image decoding apparatus, image encoding method and image decoding method |
| CN103385004B (en) * | 2011-06-30 | 2016-12-28 | 三菱电机株式会社 | Picture coding device, picture decoding apparatus, method for encoding images and picture decoding method |
Also Published As
| Publication number | Publication date |
|---|---|
| FR2809917A1 (en) | 2001-12-07 |
| EP1161099A1 (en) | 2001-12-05 |
| JP2001346217A (en) | 2001-12-14 |
| US7277487B2 (en) | 2007-10-02 |
| CN1167277C (en) | 2004-09-15 |
| US20030165194A1 (en) | 2003-09-04 |
| FR2809917B1 (en) | 2002-08-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11089311B2 (en) | Parameterization for fading compensation | |
| KR100681370B1 (en) | Predicting motion vectors for fields of forward-predicted interlaced video frames | |
| US7359558B2 (en) | Spatial scalable compression | |
| RU2409005C2 (en) | Method of scalable coding and decoding of video signal | |
| KR100849965B1 (en) | Method and apparatus for providing prediction mode fine granularity scalability | |
| JP4769605B2 (en) | Video coding apparatus and method | |
| US20060133475A1 (en) | Video coding | |
| US20040240549A1 (en) | Method and/or apparatus for reducing the complexity of H.264 B-frame encoding using selective reconstruction | |
| CN1633812A (en) | Global motion compensation for video pictures | |
| JP2008500760A (en) | Method and apparatus for encoding digital video data | |
| KR20120028843A (en) | Method and apparatus of layered encoding/decoding a picture | |
| JP2006025429A (en) | Coding method and circuit device for executing this method | |
| CN1167277C (en) | Image code having movement predicting section | |
| CN110636302A (en) | Video decoding method, video encoding method, video decoding apparatus, video encoding apparatus, storage medium, video decoder, and video encoder | |
| JP2010239423A (en) | Photographing resolution predictive video encoding and decoding apparatus | |
| JP2010050911A (en) | Encoding apparatus | |
| CN110636288A (en) | Video decoding method, video encoding method, video decoding device, video encoding device and electronic equipment | |
| CN110572675B (en) | Video decoding and encoding methods and devices, storage medium, decoder and encoder | |
| KR0172902B1 (en) | Mpeg encoder | |
| JP5298487B2 (en) | Image encoding device, image decoding device, and image encoding method | |
| CN115314712A (en) | Video encoding method using multiple loop filters in low-delay video encoding and related video encoding device | |
| JP2004007736A (en) | Device and method for decoding image | |
| KR970009379A (en) | Motion compensation device for ultra low speed transmission | |
| KR20000050508A (en) | Apparatus for hierarchial motion estimation in image coder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20040915 Termination date: 20190530 |